On Wed, Mar 08, 2017 at 03:40:52PM +0200, Mika Kuoppala wrote: > Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> writes: > > > Currently, we sum the render and media cycles (on different engines) to > > compute a percentage - but we fail to factor in the duplication into the > > threshold calculations. This makes us very eager to upclock! > > > > I wonder if this was intentional. However acting on behalf > of the more busy engine on the calculation smells more correct. Yeah, iirc it was intentional, based on the thinking that the engines share memory bw so compete against each other. However, such contention also shows up as busy cycles in both, so that seems flawed. I can't recall any other intention. -Chris -- Chris Wilson, Intel Open Source Technology Centre _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx